Hello Manned,
I have come across those issues frequently in large models, and I find the FLOODLIMIT issues are generally a result of Pump Station flows under "Start-up" conditions of the model. If it's possible, try running the model using several hours simulation time under low flows with a very small time step (1second if possible). Use that result file as a HOTSTART file.
If that doesn't resolve your issues, you may be able to change the pump station parameters away from a pump curve to a defined flow to create the HOTSTART file, and then swap it out for your actual simulation. I don't remember if changing the operational parameters of a simulation is considered too significant of a change for a HOTSTART file to be used, but I think you'll be fine.
As Ricardo has said, having a FLOOD greater then 10m is not likely accurate. If you increase the FLOODLIMIT to say, 100m, it'll still fail. You have some instability in your model that hotstarts generally resolve.